menekis/OCRmyPDF
Public Access
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
2,764 Commits 5 Branches 303 Tags
b83d7f6d1aa9c9cc27697a1e05e7d9573aefda84
Commit Graph
797 Commits
Author SHA1 Message Date
James R. Barlow b83d7f6d1a subprocess: refactor and add run_polling_stderr 2020-11-29 14:36:03 -08:00
James R. Barlow 80e957908a tesseract: fix run call with logs_errors_to_stdout 2020-11-29 14:25:46 -08:00
James R. Barlow 43f41863fa check_pdf: document how we handle linearization 2020-11-18 11:54:07 -08:00
James R. Barlow d71e50e83d Fix "readLinearizationData for file that is not linearized" 
pikepdf 2.1.0 throws wrong type of exception in this case, so special-case it.

Closes #680
Closes #681
 2020-11-18 11:52:17 -08:00
James R. Barlow 1f598da3c1 ghostscript: better docs and comments 2020-11-18 11:34:17 -08:00
James R. Barlow 5c56f61209 unpaper: type hints 2020-11-11 02:59:37 -08:00
James R. Barlow 895fddd85e Replace most uses of universal_newlines with text 
The parameters are equivalent but the latter is better named. Since
Python 3.6 doesn't support text= we use our wrapper to add it in that
place.

This is for subprocess.run.
 2020-11-07 00:48:08 -08:00
James R. Barlow 5a59e4d543 unpaper: don't use universal_newlines=True 
There's no specific reason to do this. We can log binary output equally
 well.
 2020-11-07 00:18:27 -08:00
James R. Barlow 6d3f9ff15a api: rework ocr() slightly to simplify variable handling 2020-11-03 17:10:52 -08:00
James R. Barlow b913e5dfef ghostscript: don't repeat log in debug 
Subprocess already does this for us.
 2020-11-03 01:45:06 -08:00
James R. Barlow dd8a5a4c72 Fix log domain names 
ocrmypdf.subprocess.subprocess.ghostscript -> ocrmypdf.subprocess.ghostscript
 2020-11-03 01:44:35 -08:00
James R. Barlow 36e9a54f02 Remove extraneous page rotation 
This was added in commit b5ccbfd but seems to have been ill-advised.
 2020-11-03 01:34:28 -08:00
James R. Barlow 3707af3b74 Change pdf.root to pdf.Root 2020-11-03 01:30:31 -08:00
James R. Barlow ced7ad9164 unpaper: round off DPI 2020-11-03 01:14:57 -08:00
James R. Barlow 54bbbfdeb3 Fix UnboundLocalError when considering ImageMasks for optimization 
Uncovered by test file in issue 667, although unrelated to that issue.
 2020-11-03 01:08:14 -08:00
James R. Barlow 19bf3aeb00 api: improve typing 2020-11-02 02:33:34 -08:00
James R. Barlow e86be0031c unpaper: fix process output handling 
With the ocrmypdf.subprocess wrapper, logging the output here
is redundant and loses the page number context.
 2020-11-02 01:07:41 -08:00
James R. Barlow 6425977998 unpaper: use pnm instead of png 
Some users reported problems with PNG recently; try PNM.

Fixes #665
Fixes #667
 2020-11-02 01:05:56 -08:00
James R. Barlow d57df2d980 subprocess: support programs that write their messages to stdout 2020-11-02 01:00:59 -08:00
James R. Barlow 664d0c7969 Document configure_debug_logging 2020-11-02 00:59:00 -08:00
James R. Barlow a354663ee1 Fix typo in API documentation 2020-11-02 00:58:28 -08:00
James R. Barlow d55e673d9c Fix warning about --pdfa-image-compression argument at wrong times 
Closes #663
 2020-10-27 23:09:45 -07:00
Edward BettsandGitHub 2def7e3392 Use % for percentage in string format (#643) 2020-10-27 23:09:14 -07:00
James R. Barlow 5ba56adb53 Fix page rotation issue (again) 
Commit 1327ab3 introduced a fix for a regression, which was reported
in #581, #634. It appears that the actual cause of this issue was
default parameters to rasterize_pdf_page in pluggy not working as
expected, causing a default rotation=0 even when a rotation was needed.
As such the OCR image was generated with the wrong orientation,
causing the initial regression and fix in commit 1327ab3.

Now that the real problem is identified, it's apparent that the logic
prior to 1327ab3 was found and we can revert to 1327ab3 since it fixes
all known cases including #658.

This reverts 1327ab3 except for retaining improves to rotation output.
 2020-10-24 02:45:21 -07:00
James R. Barlow b5ccbfdf25 Fix hookspec of rasterize_pdf_page to remove default parameters 2020-10-24 02:35:18 -07:00
James R. Barlow 8c35d6e6e4 Fix debug log messages being suppressed from child processes 2020-10-22 02:20:06 -07:00
James R. Barlow d1e0c81eda Ensure worker_pdf is closed after gathering info in a thread 
This is hacky, uses global state, but it does improve the situation for now.
 2020-10-22 00:38:24 -07:00
James R. Barlow 10c8e4f8b4 Only create debug.log when running from command line 
When used as a library ocrmypdf shouldn't make policy decisions, like where to
put a log file. Unsurprisingly, creating it causes problems for library users
because we deleted the temporary folder which held the log file and made no
effort to move it to a new location.

Also update the documentation to better described how an application should
handle this.

Closes #657
 2020-10-20 01:29:36 -07:00
James R. Barlow 6be2242c21 Describe "OCR" step as "Image processing" when --tesseract-timeout=0 
Fixes #647
 2020-10-08 01:03:42 -07:00
James R. Barlow 204c9d6ae1 Fix inverted colors during JBIG2 optimization on paletted images 
Fixes #640
 2020-10-07 04:08:50 -07:00
James R. Barlow 4e15eb8d14 Fix image optimization discarding image masks and soft masks associated with PNGs 
Fixes #648
 2020-10-06 03:20:54 -07:00
James R. Barlow 8b01ab8ad2 Better type checking on ocrmypdf.ocr(plugins=...) 2020-10-05 15:02:34 -07:00
James R. Barlow 4eacb3454f hOCR: write text in correct order 
Fixes #642
 2020-09-29 02:45:11 -07:00
James R. Barlow 3ef8872a1e pngquant driver: refactor, use streams instead of temporary files 2020-09-25 00:18:02 -07:00
James R. Barlow 28eec73eed Tighten unpaper-args validation to exclude . and .. 
Just in case
 2020-09-25 00:18:02 -07:00
James R. Barlow bfe4a5b329 Tidy a log message 2020-09-25 00:17:57 -07:00
Suyash BeheraandGitHub 9a6cd95e5f load zlib before liblept on windows (#633) 
fixes #631
 2020-09-17 03:14:42 -07:00
James R. Barlow d464d3122e Use img2pdf to create optimized PNG images 
Fixes #629, #620
 2020-09-17 03:11:26 -07:00
James R. Barlow 1327ab37d4 Fix page rotation regression 
Fixes #634, #581
 2020-09-17 02:57:00 -07:00
James R. Barlow 67553fc5c6 Display page numbers in log messages when grafting 2020-09-17 01:20:50 -07:00
James R. Barlow 306a903854 Remove unused function log_page_orientations 2020-09-17 01:20:02 -07:00
James R. Barlow b93cf51c0f Disable pikepdf mmap 
Infrequently we can reproduce this error:

terminating with uncaught exception of type std::runtime_error: pybind11_object_dealloc(): Tried to deallocate unregistered instance!

The error is probably related to pybind11 issue #2252 and a bunch of
other related issues. Until that is resolved in pybind11 and pikepdf
we will disable the pikepdf mmap interface.
 2020-09-16 23:48:55 -07:00
James R. Barlow 8b5b02e0d8 Expand documentation of filter_page_image 2020-09-14 14:36:17 -07:00
James R. Barlow 31994258fb metadata fixup: don't try to update original PDF's metadata with docinfo 2020-09-08 02:35:16 -07:00
James R. Barlow 1f15ecbca5 Add "Postprocessing" message as a hint for long Ghostscript runs 2020-09-08 02:34:10 -07:00
James R. Barlow e6a7b58863 Merge branch 'de-gpl' 2020-08-12 12:20:38 -07:00
James R. Barlow 9b641055e1 Fix KeyError: 'dpi' when using --threshold on image to PDF 
Fixes #607
 2020-08-07 02:21:02 -07:00
James R. Barlow 8c90f7c972 Replace GPLv3-derived PDF/A template with PostScript generator 2020-08-05 01:30:45 -07:00
James R. Barlow aa0ec40102 Change license of all GPLv3 files to MPL-2.0 
https://github.com/jbarlow83/OCRmyPDF/issues/600
 2020-08-05 00:44:42 -07:00
James R. Barlow 4cc0dc6b4a Additional size increase reasons 2020-08-03 16:03:29 -07:00